HAL Id: hal-03033204
https://hal.archives-ouvertes.fr/hal-03033204
Submitted on 4 Jan 2021
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of
sci-entific research documents, whether they are
pub-lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Polymerization Photoinitiators with Near-Resonance
Enhanced Two-Photon Absorption Cross-Section:
Toward High-Resolution Photoresist with Improved
Sensitivity
Caroline Arnoux, Tatsuki Konishi, Emma van Elslande, Eric-Aimé
Poutougnigni, Jean-Christophe Mulatier, Lhoussain Khrouz, Christophe
Bucher, Elise Dumont, Kenji Kamada, Chantal Andraud, et al.
To cite this version:
Caroline Arnoux, Tatsuki Konishi, Emma van Elslande, Eric-Aimé Poutougnigni, Jean-Christophe
Mulatier, et al.. Polymerization Photoinitiators with Near-Resonance Enhanced Two-Photon
Absorp-tion Cross-SecAbsorp-tion: Toward High-ResoluAbsorp-tion Photoresist with Improved Sensitivity. Macromolecules,
American Chemical Society, 2020, 53 (21), pp.9264-9278. �10.1021/acs.macromol.0c01518�.
�hal-03033204�
Polymerization Photoinitiators with Near-Resonance
En-hanced Two-Photon Absorption Cross-Section: Towards
High-Resolution Photoresists with Improved Sensitivity
Caroline Arnoux,
aTatsuki Konishi,
b,cEmma Van Elslande,
aEric-Aimé Poutougnigni,
aJean-Christophe Mulatier,
aLhoussain Khrouz,
aChristophe Bucher,
aElise Dumont,
aKenji
Kamada,
b,cChantal Andraud,
aPatrice Baldeck,
aAkos Banyasz,
aCyrille Monnereau
a*
a Univ. Lyon, ENS Lyon, CNRS, Université Lyon 1, Laboratoire de Chimie, UMR 5182, 46 Allée d’Italie, 69364Lyon, France. Mail : cyrille.monnereau@ens-lyon.fr (corresponding author)
b Nanomaterials Research Institute (NMRI), National Institute of Advanced Industrial Science and Technology (AIST), 1-8-31 Midorigaoka, Ikeda, Osaka, 563-8577 Japan
c Department of Chemistry, School of Science and Technology Kwansei Gakuin University, Sanda 669-1337, Japan
ABSTRACT: We take advantage of the near-resonant enhancement of the third-order nonlinear response to
engi-neer a two-photon polymerization photoinitiator with optimized efficiency in regard with literature benchmarks. We study in detail its linear and resonant third-order nonlinear optical properties, with a particular focus on photoinduced radical generation. Careful choice of peripheral substituents enable its direct solubilization into a mixture of com-mercial multifunctional acrylic monomers, resulting in homogeneous photoresist with good optical transparency upon simple mixing. When submitted to sub-nanosecond pulsed laser irradiation at 532 nm, the resulting photore-sist displays a polymerization threshold up to eight times lower than that obtained with classically used two-photon initiators, giving rise to sub-100 nm linewidth at 250 nm interline separation, and resulting in highly defined three-dimensional (3D) structures.
INTRODUCTION
Within the last decade, the additive bulk manufacturing of 3D structures has become a major driving force in modern industry and is generally considered as one of the nine
pillars of the so-called “4th industrial revolution”, or
indus-try 4.0.1-2
Once limited to rapid prototyping of industrial high-value components and models, nowadays 3D printing has in-deed become a relatively standard and affordable tech-nology that is used in the manufacturing of many daily-life
goods from plastic housing-shell to medical prostheses.
3-5 In the near future, 3D printing, owing to its unlimited
de-sign flexibility, is expected to substitute polymer ther-moforming or milling in high-tonnage productions such as automotive and aerospace industries or the construction sector.6-8
As the range of applications that additive 3D manufactur-ing encompasses constantly expands, more and more de-manding specific technical requirements and constraints
tend to emerge.9-10 In particular, fabrication of nano and
micro-structured objects require resolution that are not
ac-cessible with standard layer-by-layer photolithography.
11-13 In that framework two-photon absorption-induced
ste-reolithography is generally considered as a potential way
to circumvent the limitations in achievable spatial
resolu-tion.14-15 However, the intrinsic characteristics of
two-pho-ton absorption, which allows localized polymerization re-action in the volume of the focal point of the laser beam, makes it very difficult to extend the polymerization to the
fabrication of large structures with such high resolution.
16-17
As a matter of fact, to become more than a “laboratory curiosity”, two-photon additive manufacturing needs to face the challenge of high-throughput, large area printing
of surfaces.17-24 To this end, endeavors are focused on the
optimization of both optical setups16, 25-26 and photoresist
sensitivity.
From a chemical point of view, the design of an improved photoresist relies on a combination of a reactive monomer with high functionality, capable of generating an intricate
polymer network even at low monomer conversion,27 and
a highly sensitive photoinitiator (PI).28 Monomer sensitivity
can be improved by removing inhibiting species such as molecular oxygen or exogenous inhibitors (such as MEHQ) generally included as stabilizers in monomer
for-mulations.29 However, such strategies can severely
im-pair the accessible resolution, as they favor radical diffu-sion outside of the irradiated volume, and result in a loss
of control on voxel volume.30-31 Therefore, a majority of
re-search works have been focused on improving PI effi-ciency.32-34
2
As a general statement, PI efficiency essentially relies on two parameters, namely i/ the two-photon absorption
cross-section (σTPA), which quantifies the probability for
the chromophore to undergo transition to its excited state upon laser irradiation through the simultaneous absorp-tion of two photons, ii/ the radical generaabsorp-tion quantum ef-ficiency, which represents the probability for the excited molecule to generate a radical capable of initiating the
re-action of polymerization.35
While a lot of recent researches have focused on the use of PIs with optimized radical generation quantum
effi-ciency,34, 36-41 the improvement of σ
TPA has exclusively
been addressed through a relatively classical
chromo-phore engineering.42-50 This followed the Equation 1 that
approximates in the so-called “three-level model” the im-aginary part of the third-order hyperpolarizability γ, re-sponsible for the two-photon absorption phenomenon, where P denotes a permutation operator over the optical
frequencies; 𝜇𝑖𝑗, the transition dipole moment between
the singlet state i and the singlet state j; 𝐸𝑖𝑗, the transition
energy; Δ𝜇𝑖𝑗,the dipole moment difference between the i
and j states and Γ the associated damping factor; ℏ𝜔, en-ergy of the incident photon.
In all of these works, a general paradigm is that dipolar PIs should be selected so that their one-photon absorp-tion (OPA) resonance wavelength lies closely to half of
that of the operating laser51 By doing so, the sum of the
energies of two incident photons is resonant with the en-ergy gap between the ground and excited states of the PI
(ie the S1 state in the case of a dipole, and the S2 state in
the case of a quadrupole). Regarding the lowest TPA en-ergy transition, this strategy leads to a minimization of the second term of the denominator for the D (dipolar mole-cules) or T (quadrupolar molemole-cules) terms, thus to a
max-imization of the γ term. 52
Im𝛾(−𝜔; 𝜔, −𝜔, 𝜔) = Im 𝑷 [ + 𝜇01 2 Δ𝜇 01 2 (𝐸01− ℏ𝜔−𝑖Γ)2(𝐸01− 2ℏ𝜔−𝑖Γ) (𝐷) + 𝜇012 𝜇122 (𝐸01− ℏ𝜔−𝑖Γ)2(𝐸02− 2ℏ𝜔−𝑖Γ) (𝑇) − 𝜇014 (𝐸01− ℏ𝜔−𝑖Γ)2(𝐸01+ ℏ𝜔+𝑖Γ) (𝑁) ] (1)
However, another way of maximizing γconsists in
mini-mizing the denominator terms by using an incident photon
that comes close to the OPA resonance ℏ𝜔 ≈ 𝐸01 yet
does not overlap with the linear absorption band This phenomenon is known as the near-resonance
en-hancement of the σTPA,53-55 and has been shown to play a
significant role in the outstanding σTPA values obtained for
molecules such as extended porphyrin56 or squaraine57
derivatives mainly used in two-photon photodynamic
ther-apy.58 To the best of our knowledge, only one example
has been explicitly reported to date, that concerns two-photon induced microfabrication using cationic polymeri-zation.59
Practically, working with a 532 nm pulsed laser, a blue absorbing chromophore structurally optimized following the general principles and guidelines of chromophores design for nonlinear optics (large π-conjugated structure, presence of electron donating and/or withdrawing groups)
can be used as a near-resonant two-photon photosensi-tizer.
In this paper, we describe the synthesis and detailed char-acterizations of two such chromophores, using a common set of monomeric building blocks. These chromophores, named V-Shape and Oct, consist respectively in two and three branched oligo(phenyleneethynylene) derivatives, end terminated with aniline moieties used as strong elec-tron donors. We investigate in detail their linear and non-linear spectroscopic features, and study their mecha-nisms of decomposition upon photoirradiation. By com-parison with commercially available benchmark Norrish I (Irgacure OXE2) and II (BDEBP and ITX) type of PIs (as shown in Figure 1), which have been previously used in
two-photon microfabrication, 60-65 we identify the nature of
the photogenerated radicals and propose a mechanism accounting for their formation.
Figure 1. Chemical structures of the five PIs investigated in this study and the two monomers that were used: (a)
V-Shape; (b) Oct; (c) 4,4′-Bis(diethylamino)benzophenone also called Michler’s ethyl ketone BDEBP; (d) ITX; (e) OXE2 (f) 1,10-decanediol diacrylate DDA; (g) dipentaerythritol penta/hexa-acrylate DPPHA.
3
Through a combined spectroscopic and electrochemical study, we study in details the mechanisms of radical gen-eration and polymerization initiation. Finally, we show that the two new PIs reported in the study, when incorporated in a mixture of multifunctional acrylic monomers, provide significant benefits in terms of polymerization sensitivity with a polymerization threshold lowered up to eight times in comparison with archetypical Norrish I (OXE2, acylox-ime ester) or Norrish II (Michler’s ketone derivative, 4,4′-bis(diethylamino)benzophenone BDEBP and isopropyl thioxanthone ITX) PIs similar to those used in recent works dealing with two-photon microfabrication, while maintaining high-resolution features.
EXPERIMENTAL SECTION
Materials and general methods. All reactions were
performed under argon. Solvents were used without fur-ther purification. All reagents, including reactant 2 were purchased from Sigma Aldrich, TCI or Alfa Aesar and were used without further purification. The reaction pro-gresses were monitored by thin layer chromatography (TCL) with silica gel 60 F254 using UV lamp detection. Column chromatography was performed using Acros Or-ganics (0.035-0.70 mm) silicagel for purification. Dipen-taerythritol penta-/hexa-acrylate (DPPHA) was purchased from Sigma Aldrich and 1,10-decanediol diacrylate (DDA) from TCI. 4,4′-bis(diethylamino)benzophenone also called Michler’s ethyl ketone BDEBP was purchased from
Merck and OXE2 from Angene Chemical. 1H and 13C
NMR spectra in CDCl3 were recorded at room
tempera-ture on a BRUKER Avance spectrometer operating at 400
MHz for 1H and 75 MHz for 13C respectively. Chemical
shifts are reported in parts per million (δ, ppm) relatively to tetramethylsilane, using residual solvent peaks as
in-ternal standards (CHCl31H: 7.26 ppm, 13C: 77.36 ppm).
Spectra are provided in supporting information (see Fig-ures S1 and S2). High resolution mass spectrometry mea-surements were performed at Centre Régional de Me-sures Physiques de l'Ouest (Université de Rennes, Rennes, France).
Synthesis. Synthesis of compounds 1, 3 and 4 have been
already reported in the literature ( ). 66-67
V-Shape: 1 (465 mg, 1.21 mmol) and 2 (114 mg, 0.36
mmol) were dissolved in a mixture of THF (20 mL) and
Et3N (15 mL). The solution was degassed by argon
bub-bling for 1 hour. Then Pd(PPh3)2Cl2 (138 mg, 5.4%) and
CuI (69 mg, 10%) were added. The reaction mixture was stirred at 90°C under argon for 48 hours. Solvents were removed under reduced pressure and the crude product was dissolved in dichloromethane. The solution was
washed with NH4Cl, water and pre-dried with saturated
NaCl solution. The solution was dried with Na2SO4,
fil-tered and solvents were remove under vacuum pressure. The product was then purified on silica gel with a mixture
PE/CH2Cl2 as eluent (from 9/1 to 8/2, v/v. solid deposition
of the product) to give a brown oil (195 mg, 58%). 1H NMR
(400 MHz, CDCl3): δ 7.65 (d, J = 1.4 Hz, 2H), 7.64 (t, J = 1.3 Hz, 1H), 7.49 (s, 8H), 7.39 (d, J = 8.9 Hz, 4H), 6.60 (d, J = 9.0 Hz, 4H), 3.30 (t, J = 8.1 Hz, 6H), 1.61 (t, J = 7.0 Hz, 8H), 1.36 – 1.33 (m, 24H), 0.93 (t, J = 6.7 Hz, 12H). 13C NMR (75 MHz, CDCl 3): δ 148.2, 133.9, 133.0, 132.3, 131.6, 131.2, 125.3, 124.9, 122.0, 121.1, 111.2, 108.2, 93.5, 91.3, 88.5, 87.0, 51.0, 31.7, 27.2, 26.8, 22.7, 14.1.
HRMS (ESI) m/z calcd for [C62H71BrN2 + H]+ at 923.48734,
found 923.4871.
Oct: 3 (700 mg, 1.44 mmol) and 4 (65 mg, 0.43 mmol)
were dissolved in a mixture of THF (15 mL) and Et3N (10
mL). The solution was degassed by argon bubbling for 1
hour. Then Pd(PPh3)2Cl2 (150 mg, 5%) and CuI (80 mg,
9.6%) were added. The reaction mixture was stirred at 70°C under argon for 24 hours. Solvents were removed under reduced pressure and the crude product was dis-solved in dichloromethane. The solution was washed with
NH4Cl, water and pre-dried with saturated NaCl solution.
The solution was dried with Na2SO4, filtered and solvents
were remove under vacuum pressure. The product was
then purified on silica gel with a mixture PE/CH2Cl2 as
el-uent (from 9/1 to 8/2, v/v. solid deposition) to give a yellow
oil (150 mg, 28%). 1H NMR (400 MHz, CDCl 3): δ 7.64 (s, 3H), 7.48 (s, 12H), 7.37 (d, J = 8.8 Hz, 6H), 6.58 (d, J = 9.0 Hz, 6H), 3.28 (t, J = 7.4 Hz, 12H), 1.59 (t, J = 7.7 Hz, 12H), 1.41 – 1.27 (m, 36H), 0.91 (t, J = 6.6 Hz, 18H). 13C NMR (75 MHz, CDCl3): δ 148.1, 134.0, 133.0, 131. 6, 131.2, 124.7, 124.1, 121.4, 111.2, 108.3, 93.4, 90.6, 89.2, 87.0, 51.0, 31.7, 27.2, 26.8, 22.7, 14.1. HRMS (ESI) m/z
calcd for [C90H105N3 + H] + at 1228.83813, found
1228.8375.
Absorption and luminescence. Absorption spectra
were recorded on a JASCO V-650 spectrometer in dilute
dichloromethane solution (about 10-6 mol/L). Emission
spectra were measured using a Horiba-Jobin-Yvon Fluorolog-3 spectrofluorimeter equipped with a three-slit double grating excitation and emission monochromator with a dispersion of 2.1 nm/mm (1200 groove/mm). A R928 detector was used. Corrections
were applied for both the excitation light intensity variation (lamp and grating) and the emission spectral response (detector and grating). Absolute fluorescence quantum yields were measured in diluted dichloromethane solu-tions using a GMP G8 integrating sphere. Sample were introduced into a capillary quartz tube placed into the cav-ity of the sphere. Differences between the peak intensities of the lamp and the emission profiles were minimized us-ing a density filter (0.5%) to attenuate the lamp signal and thus ensure that the collected signal remained in the lin-ear range of the detector.
Z-Scan TPA measurement. Z-scan measurements were
conducted in dichloromethane (spectroscopic grade) so-lutions for V-Shape (0.61 mM), Oct (1.52 mM) and
BDEBP (9.8 mM). The open aperture z-scan method 68
was employed for all samples held in 2-mm quartz cu-vette.
4
Scheme 1. Synthesis of target compounds V-Shape and Oct.
Two measurement procedures were used: (1) wavelength (WL)-scan measurement, where the single-power meas-urement was repeated for difference excitation wave-lengths and the z-scan traces were analyzed by assuming TPA process; and (2) power-scan measurement, where measurements were repeated at least with four different incident powers for each excitation wavelength to confirm that the observed nonlinear absorption is dominated by TPA. The combination of these measurement procedures allows us to obtain fast and reliable results because the WL-scan results show the overview of the spectra and the power-scan results for some selected wavelengths make sure that the observed results are caused by TPA. The incident powers (excitation intensities) used differed de-pending on the wavelength range of the laser i.e. the con-figuration of the femtosecond optical parametric amplifier
Spectra-Physics TOPAS Prime and were 0.40 mW
(80~100 GW/cm2) for 600~900 nm range (SHS, SHI) and
0.25 mW (60~170 GW/cm2) for 485~600 nm range (SFS,
SFI) for the WL-scan. On the other hand, power-scan was made for these powers and three weaker powers (factors of 0.7, 0.5 and 0.3). The typical pulse width and the Ray-leigh range were 90~120 fs and 6~11 mm although they
changed depending on the wavelength. MPPBT/DMSO 68
and bis-MSB /CLF 69 were used as reference sample for
600–900 nm and 485–600 nm respectively.
EPR-Spin Trapping (ST). Dichloromethane solutions of
each PI in presence of trapping agent (N-tert-Butyl-α-phe-nylnitrone PBN, and 5,5-Dimethyl-1-pyrroline N-oxide DMPO) were carefully degassed through five freeze-pomp-thaw cycles in quartz EPR tubes. The spectra of the irradiated samples (M365L2 Thorlabs LED, 365nm) were recorded in situ. EPR assays were all carried out at room temperature using a Bruker E500 spectrometer operating at X-band (9.33-9.87 GHz), standard cavity, with 100 KHz modulation frequency. The instrument settings were as the following: microwave power: 0.2-64 mW; modulation amplitude: 1 G. Hyperfine coupling constants a values were obtained with simulation of experimental spectra us-ing Easyspin (Matlab toolbox) and compared with the lit-erature.
Cyclic Voltamperometry. Cyclic voltammetry (CV)
curves have been recorded using a SP300 Biologic po-tentiostat. All studies were conducted under an argon
at-mosphere in a standard one-compartment three-elec-trode electrochemical cell. An automatic ohmic drop com-pensation procedure was systematically performed when using CV. Vitreous carbon (Ø = 3 mm) working electrodes (CH Instruments) were polished with 2 µm diamond paste
(PRESI SA) before each recording. Ag/AgNO3 (CH
Instru-ments, 10−2 mol/L + TBAP 10−1 mol/L in CH3CN) was used
as a reference electrode and bis(η5-cyclopentadienyl)iron
(ferrocene) was used as an internal reference. Dichloro-methane (99.9 %, extra-dry, stabilized with amylene) was purchased from Acros and used as received. Tetra-n-bu-tylammonium perchlorate (TBAP, Fluka puriss.) was pur-chased and dried under vacuum at 60°C for 24 h before use.
DFT calculations. Density functional theory (DFT)
calcu-lations were performed with the Gaussian 16 Rev B.01
suite of programs.70 Structures of V-Shape in the neutral,
radical cationic and radical anionic states were computed at the M06HF/6-31G(d,p) level of theory, both in gas phase and in dichloromethane to estimate electronic en-ergies as well as redox potentials. Absolute redox poten-tials were estimated at 298K, after frequencies calcula-tions according to the computational protocol proposed by
Truhlar and coworkers.71
Stern-Volmer experiments. Emission spectra of the
tested PIs in dichloromethane were measured using a Horiba-Jobin Yvon Fluorolog-3 spectrofluorimeter from 375 to 700 nm at room temperature with an excitation at
λex = 370 nm. A concentrated solution of DPPHA at
iso-molar concentration of PI as in the first solution was then added until reaching the final concentration of DPPHA. No change was observed, indicating an absence of electron or energy transfer between the singlet excited state of
V-Shape or Oct and the monomer.
Photolysis studies. Photolysis studies were performed
in a home-made setup consisting of a 365 nm Thorlabs LED (M365L2) equipped with a convex lens (plano-con-vex f = 25.4 mm, Thorlabs) aiming to obtain a beam diam-eter of about 9 mm through the sample cuvette (10x10mm) placed 10 cm downstream of the LED. LED power was measured on a thermal power sensor (S405C, Thorlabs) placed at the same distance. Prior to the study, dichloromethane solutions of PI in presence or absence of monomer were carefully degassed through five freeze-pomp-thaw cycles and were stirred during photolysis.
5
Microfabrication. Photoresist preparation: the PIs were
mixed with DDA and stirred for 30 min at room tempera-ture with additional dichloromethane when needed. No solvent was necessary for V-Shape and Oct solubiliza-tion. DPPHA was then added. The solution was stirred mechanically for 1 min and then magnetically for 30 min. Good homogeneity was obtained without filtration, ensur-ing that the desired concentration was achieved in each photoresists tested.
Fabrication and setup: the 3D microfabrication was
car-ried out using a Microlight3D printer µFAB-3D based on a Zeiss Axiovert 200 inverted microscope equipped with a XYZ piezo nanomanipulator allowing the translation of the sample relative to the laser focal point, and a CMOS cam-era mounted behind a dichroic mirror for monitoring of the polymerization process. The laser module includes a mi-crochip self-Q-switched frequency-doubled Nd:YAG laser at 532 nm providing 560 ps pulses at a repetition rate of 11.7 kHz with a maximum average power of 11.5 mW at entrance of the objective. Average laser power were measured at the entrance pupil of the objective on a standard photodiode power sensor (S120VC, Thorlabs). The incident beam was focused with a high numerical ap-erture objective (x100, NA 1.40, oil immersion, Zeiss Plan-APOCHROMAT). The laser power was controlled by an acousto-optic modulator which was computer-piloted via LITHOS software as well as the displacement of the sam-ple and the scanning speed. The samsam-ple consists of a drop of photoresist deposited on a borosilicate coverslip (170 µm ± 5 µm thick). After the manufacturing process, the structures were obtained by washing away the unre-acted monomer with two successive 10-minute acetone baths.
Microstructure characterization: samples were metallized
with 5 nm platinum via vacuum deposition with Leica EM ACE600 sputter coater to prevent charge accumulation and enhance the contrast during the following Scanning Electron Microscopy (SEM) observations. The line widths were measured using a top view and the line heights us-ing a grazus-ing view (90°) with a Zeiss Supra 55VP SEM. A home-made python program was used to automatically determine the edges of the lines and give an average value of a line width over a length of several micrometers of the middle of each line. More details are provided in
supporting information. The error bars correspond to the 95% confidence interval given by the student's
t-distribu-tion. Thus, it provides information onthe error resulting
from the measurement of the line widths and heights by SEM. The values were considered without subtracting the thickness of the platinum layer (expected at twice 5 nm
i.e. 10 nm).
RESULTS AND DISCUSSION
Synthesis. V-Shape and Oct were synthesized using a
general synthetic approach and building blocks depicted
in previous articles.66-67 The final key-step of both
synthe-ses consists in a Sonogashira coupling of respectively a tribromobenzene 2 with the extended alkyne terminated aniline 1, and three iodoalkyne chains 3 with a 1,3,5-tri-ethynylbenzene 4 (see
), which synthesis has already been reported.66
Depend-ing on the reaction conditions, it is possible to drive the reaction towards the almost exclusive formation of either
V-Shape or Oct. The preferential formation of V-Shape
rather than of Oct following the first pathway could be ex-plained by the electron enrichment of the halide through the first two couplings with the alkyne-aniline moiety, which is detrimental to the oxidative addition step of Pd(0) into the remaining carbon-halide bond. Both target mate-rials can be recovered in moderate yields after purification of the crude product on silica gel. Detailed synthetic pro-tocols and characterizations are featured in the experi-mental section.
Spectroscopy. Main spectroscopic data for both studied
molecules are gathered in Erreur ! Source du renvoi
in-trouvable., along with those of the studied benchmark
PIs, when available. Both molecules are characterized by a broad and structureless absorption band (see Figure 2) that peaks at the limit between near UV and visible (390 nm) and extends into the violet and blue range of the vis-ible spectra. The shape of the spectrum in dichloro-methane solvent, along with the characteristic large oscil-lator strength, is typical of an intramolecular charge trans-fer (ICT) band, characteristic of two-photon absorbing chromophores. In addition, a similarly broad fluorescence band is observed peaking at 508 nm and 495 nm for
V-Shape and Oct respectively.
Table 1. Summary of the main photophysical and electrochemical parameters for both molecules and references in dichloro-methane: maximum absorption wavelengths (λ max-abs), molar extinction coefficients (ε) at λmax-abs, maximum emission wave-lengths (λmax-em), fluorescence quantum yields φf, cathodic peak potential (Epc), anodic peak potential (Epa), standard free energy changes for photoinduced electron transfer (𝜟𝑮𝑬𝑻𝟎 ) and TPA cross-sections at 532 nm (𝝈𝑻𝑷𝑨𝟓𝟑𝟐𝒏𝒎).
PI λ(nm) max-abs ε (104 L/mol/cm) λmax-em
(nm) φf (%) Epc (V) Epa (V) ΔG𝐸𝑇 0 (kJ/mol) σ532𝑛𝑚𝑇𝑃𝐴 (GM) V-Shape 386 8.9 508 66 -2.33 +0.37 -11.4 1500±200 Oct 386 14.8 495 79 -2.48 +0.36 -2.3 740±150 BDEBP 362 4.1 - - na +0.49 na 77±11 ITX 386a 0.7a 420a 9a na na na na OXE2 340 2.1 464 ≈ 0 na na na na
6
Peak potential values measured by CV for V-Shape,OctandBDEBP (1 mmol/L) in the presence of tetra-n-butylammonium perchlorate (0.1 mol/L), E(V) vs. Eref[Fc/Fc+], at a vitreous carbon working electrode Ø = 3 mm, 298 K, ν = 0.1 V/s. na: not available. a previously re-ported,72 in chloroform.
TPA spectra were measured for both compounds by the Z-scan technique, over a range of wavelengths covering the 500-1000 nm region. In both cases, in addition to the
“regular” TPA band corresponding to the S0→S1 transition
(excitation to the lowest lying excited state, similar to that reached upon OPA at 390 nm), another strong TPA tran-sition is also seen, as expected, as the laser wavelength gets closer to the OPA resonance wavelength.
Interest-ingly while σTPA peaks ca. 500 GM for the regular TPA
band, consistent with values usually found on related
chromophore with similar π-conjugated architectures, 66,
73-74 a much higher value is found at the near-resonant
TPA band. In particular, at the operating wavelength of the frequency doubled Nd-YAG laser used in our
micro-fabrication setup (532 nm), σTPA reaches a value about
1500±100 GM for V-Shape, then increases up to a value
ca. 2000 GM at 500 nm. The 532 nm value of Oct is
sig-nificantly lower (740±150 GM), because of a more ener-getic onset of the near-resonant transition, consistent with the slightly blue-shifted onset of the ICT band in the OPA spectrum, but reaches a similar value at 500 nm (2200
GM). These very high σTPA values are reminiscent of those
found for expanded porphyrin and squaraine derivatives, for which near-resonant enhancement of TPA has been shown to play an important role in the nonlinear re-sponse.56-57, 75
Note that the quadratic dependence of the absorption pro-cess to the average laser intensity was checked at differ-ent characteristic wavelengths (denoted as “power scan” in Figure 2 and Figure S3, and including 532 nm, Figure S4) to ascertain the two-photon nature of the monitored
process at these wavelengths.76
EPR spin trapping (EPR-ST) studies. To further
demon-strate the applicability of our newly designed chromo-phores for two-photon microfabrication, photoinduced radical generation upon continuous irradiation of dichloro-methane solution of V-Shape was studied by means of EPR-ST spectroscopy. To facilitate identification of the generated radical, EPR signals and spectral simulations were systematically compared to those obtained with lit-erature benchmarks. In addition to our aforementioned Norrish I (OXE2) and Norrish II (BDEBP) PIs, azobisiso-butyronitrile (AIBN) was also used to verify the reliability of the approach. AIBN is known to efficiently produce a single radical species upon thermal or photodegradation, and should in principle lead to a unique EPR signal. More-over, in order to facilitate radical detection, radical scav-engers N-tert-butyl-α-phenylnitrone (PBN), and 5,5-dime-thyl-1-pyrroline N-oxide (DMPO) were used for each com-pound, and the results obtained with those two scaven-gers were systematically compared.
EPR spin trapping (EPR-ST) studies. To further
demon-strate the applicability of our newly designed chromo-phores for two-photon microfabrication, photoinduced radical generation upon continuous irradiation of dichloro-methane solution of V-Shape was studied by means of EPR-ST spectroscopy.
Figure 2. Normalized absorption (solid lines) and emission spectra (dashed lines) of V-Shape (red) and Oct (purple) in dichloromethane (top figure). The red-tail of OPA spectra (solid lines) along with TPA spectra of V-Shape (red) and Oct (purple) in dichloromethane obtained by wavelength (WL)-scan measurement (empty triangles) at a constant incident power and power scan measurement (full circle) performed with four irradiation powers at each wavelength, by which it proves that the observed signal originates from TPA for the selected wavelengths (bottom figure).
To facilitate identification of the generated radical, EPR signals and spectral simulations were systematically com-pared to those obtained with literature benchmarks. In ad-dition to our aforementioned Norrish I (OXE2) and Norrish II (BDEBP) PIs, azobisisobutyronitrile (AIBN) was also used to verify the reliability of the approach. AIBN is known to efficiently produce a single radical species upon thermal or photodegradation, and should in principle lead to a unique EPR signal. Moreover, in order to facilitate radical detection, radical scavengers N-tert-butyl-α-phe-nylnitrone (PBN), and 5,5-dimethyl-1-pyrroline N-oxide (DMPO) were used for each compound, and the results obtained with those two scavengers were systematically compared. 300 400 500 600 700 0.0 0.2 0.4 0.6 0.8 1.0 Norm alized Absor ba nce an d Emission Wavelength (nm) V-Shape Oct 400 500 600 700 800 900 0 1000 2000 3000 4000 5000 V-Shape Oct Wavelength (nm) Mo lar Extinction Coe fficien t ( L/m ol/cm ) 0 500 1000 1500 2000 2500 3000 TPA Cross Section (G M)
7
As expected, a single radical species was obtained in the presence of AIBN, whatever scavenger considered. The features of the trapped radical species as monitored in
ei-ther PBN (aN = 14.1, aH = 2 G) or DMPO (aN = 13.0 G, aH
= 8.9 G, aH = 1.5 G) are consistent with a radical species
originating from the classical decomposition mechanism of AIBN, after possible oxidation with traces of oxygen contained within the medium as found in previous works
with this initiator (see Figure S5).77
Irradiation of OXE2 gives a more complex pattern; in ad-dition to the previously reported spin-adduct
correspond-ing to an acetoxyl COO radical (adduct of PBN: aN = 13.5
G and aH = 1.6 G; adduct of DMPO: aN = 12.8 G, aH = 10.3
G and aH = 1.1 G),36, 78 three other species were also
iden-tified. In the first, hyperfine coupling constants are
con-sistent with the methyl radical CH3 (adduct of PBN: aN =
15.1 G and aH = 3.4 G; adduct of DMPO: aN = 14.7 G and
aH = 20.5 G), which agrees with a mechanism consisting
in a CO2 elimination from the acetoxyl radical. A last
ad-duct, which we attribute to iminyl radical resulting from the N-O band photocleavage was similarly observed with
both trapping agents (adduct of PBN: aN = 13.9 G and aH
= 2.3 G; adduct of DMPO: aN = 14.5 G and aH = 15 G). In
addition, an adduct of DMPO highlights the presence of another species consistent with reported values for acetyl radical (aN = 13.0 G, aH = 7.0 G and aH = 1.5 G) even
though the presence of this species is not fully under-stood. Note that this radical may also be present in PBN, but its expected feature may render it undistinguishable from the iminyl radical. In addition to these signals, a
tri-plet signal due to the degradation of DMPO (aN = 13.9 G)79
is visible in the corresponding EPR spectrum (see Figure S5).
In contrast, strong similarities are seen for the dominant radical species obtained with BDEBP and V-Shape (see below and Figure 3). By analogy with previous studies and general agreement on Michler’s ketone (or related
mole-cules80) photoinitiation mechanism,81 and taking in
consid-eration the structural analogies between the former and
V-Shape, we attribute the observed first radical species
trapped with PBN to a C-centered radical in alpha position
of the aliphatic chain, relative to the aniline group (aN-BDEBP
= 14 G, aH-BDEBP = 2.2 G; aN-V-Shape = 14.6 G, aH-V-Shape =
2.5 G). The hyperfine coupling constants indeed fully
match those previously described in literature.82-83 A
sec-ond signal common to both PIs (aN = 8.1 G) can be
at-tributed to a degradation of PBN, 79 finally a third species
is observed with BDEBP exclusively (aN-BDEBP = 14.8 G,
aH-BDEBP = 4.6 G): its signature corresponds to a
C-cen-tered radical, located on the carbonyl moieties, and is
typ-ically encountered with benzophenone derivatives.84-85 It
is therefore absent from V-Shape signature, the latter be-ing devoid of a carbonyl function.
By analogy with literature, the first signal observed in
DMPO (aN-BDEBP = 14.5 G, aH-BDEBP = 21 G; aN-V-Shape = 14.5
G, aH-V-Shape = 21.5 G) may correspond to adducts of the
same species described with a possible superimposition of the two C-centered signals in the case of BDEBP. Due to the better ability of DMPO to trap N and O-centered
radicals, we attribute the second and major species (a
N-BDEBP = 13.5 G, aH-BDEBP = 11.7 G; aN-V-Shape = 13.4 G, a
H-V-Shape = 11.4 G) to an N-centered (i.e. on the aniline nitro-gen) cation radical. Eventually, a signal corresponding to
the degradation of DMPO was observed for both PIs (aN
= 13.9 G).79 These species are consistent with a
mecha-nism consisting an initial electron transfer step, which then evolves through hydrogen abstraction from another PI molecule or the solvent. It constitutes another indication that V-Shape (and Oct, by extension) most probably pro-ceed through an initial homo-intermolecular electron transfer, and that the resulting radical possibly evolves by hydrogen abstraction, a point that also receives support from TD-DFT computational modeling (vide infra).
Figure 3. EPR spectra of spin trap adducts generated after irradiation of BDEBP and V-Shape at 365 nm in presence of PBN or DMPO in dichloromethane. Experimental spectra (blue) and simulated spectra (orange).
Electrochemical studies. The thermodynamic feasibility
of such a photoredox process was further demonstrated by electrochemical methods. The oxidation and reduction potentials of V-Shape and Oct were determined by CV measurements carried out in deoxygenated dichloro-methane in the presence of tetra-n-butylammonium (0.1 mol/L). As a general statement, the thermodynamic driv-ing force for electron transfer (ET) can be calculated usdriv-ing the following equation, sometimes abusively referred to
as the Rehm-Weller relation (see Equation 2):86
𝛥𝐺𝐸𝑇= 𝐹[𝐸°(𝐷+/𝐷) − 𝐸°(𝐴/𝐴−) − 𝛥𝐸00] + 𝑤𝑝 (2)
Where 𝐹 is the Faraday constant, 𝐸°(𝐷+/𝐷) and 𝐸°(𝐴/
𝐴−) stand for the donor and acceptor oxidation and
reduc-tion potentials, respectively (or the lowest oxidareduc-tion and reduction potential of the PI, in the case of an
homo-inter-molecular process like here), 𝛥𝐸00 (eV) is the so-called
zero-zero transition energy, i.e. the energy difference be-tween the vibrationally relaxed ground and excited states of the molecule, which corresponds in the case of a fluor-ophore to the crossing point of its normalized absorption
and emission spectra, and 𝑤𝑝 is the Coulomb stabilization
energy associated with the intermediate radical ion pair interaction energy with the solvent. The latter is usually neglected in a first approximation.
8
As can be seen in Figure 4, the CV curves of V-Shape and Oct display similar signatures, with one irreversible oxidation wave observed at about 0.35 V and one irre-versible reduction wave between -2.3 and -2.5 V (see Table 1). The irreversibility of these CV waves reveal the limited stability of both the oxidized and reduced species generated at the electrode interface. The cation radical moieties produced at 0.35 V on the CV curves of V-Shape and Oct was found to be unstable at all investigated scan rates (100-20 V/s) and the existence of chemical reactions coupled to the electron transfer was revealed on the CV curves by the observation of additional waves on the
re-turn scan (E -0.5 V). Both the oxidation potential and
observed behavior agree well with the formation of an
(Hex)2ArN+ cation radical.87 Density functional theory
(DFT) calculations carried out on the oxidized and re-duced forms of V-Shape suggest the formation of a stable radical anion, with an extensive delocalization of the elec-tron density contributing to its stability whereas the radical cation is localized at the aniline moieties (see energetic diagram and spin densities given in Figure S6). The latter is prone to deprotonate spontaneously with a marked ex-ergonicity at the aliphatic C adjacent to the aniline center, in good agreement with the conclusions drawn on the ba-sis of EPR data. Calculated oxidation potentials agree well the ones measured experimentally (see ESI). On the cathodic side, the reductions of V-Shape and Oct are observed at significantly different potential values.
This shift of 150 mV observed between those two
irre-versible waves is however fully consistent with the pres-ence of an electron withdrawing bromine atom on
V-Shape; which supports the conclusion that the electron
transfer is centered in both cases on the conjugated pol-yaryl backbone.
Figure 4. Cyclic voltammetry of V-Shape,OctandBDEBP (1
mmol/L) in dichloromethane in the presence of tetra-n-bu-tylammonium perchlorate (0.1 mol/L), E(V) vs. Eref [Fc/Fc+], at a vitreous carbon working electrode Ø = 3 mm, 298 K, ν = 0.1 V/s.
The distribution of spin densities obtained at the DFT/M06HF/6-31G(d,p) shown in Figure S6 indicates a stronger delocalization of the radical anion, hence more stable. Here again, the irreversibility of both CV waves re-veals the unstable character of the electrogenerated
an-ion radicals which could potentially be involved in hydro-gen abstraction, radical-radical coupling reactions or in
the hydrogenolysis of the carbon-halogen bond. 88
From the spectroscopic data, 𝛥𝐸00 is found at 440 nm and
433 nm for V-Shape and Oct respectively (2.82 and 2.87 eV). From Equation 2, a standard Gibbs energy for the ET between two V-Shape or Oct molecules can be calcu-lated, corresponding to -11.4 and -2.3 kJ/mol. Similar ex-periments were also performed on BDEBP (see Figure 4), unfortunately reduction potentials lie outside of the acces-sible electrochemical window, precluding calculation of the free energy for the homo-intermolecular electron transfer in this case.
CV measurements were also conducted on the monomer. No accessible oxidation or reduction potential could be recorded within the electrochemical window (see Figure S7), a priori discarding the possibility of a photoinduced ET with any of the PIs during the photoinitiation process,
in contrast with other literature reports.89 This was further
confirmed by a Stern-Volmer experiment90 where
evolu-tion of the luminescence intensity of V-Shape and Oct was monitored upon incremental addition of increasing monomer (DPPHA) concentration (see Figure S8). As no evolution was seen even at high monomer concentration, it can be concluded that no electron or energy transfer takes place between the singlet excited state of V-Shape or Oct and the monomer.
In summary, these data clearly outline the possibility of an exergonic photoinduced disproportionation for V-Shape and Oct. These data, in line with EPR results, strongly suggest that this electron transfer constitutes the initial step leading to radical formation and polymerization initi-ation in the microfabriciniti-ation process.
Photolysis experiments. Spectroscopic study of the
photolysis of PIs was then undertaken in order to gain more insight into the kinetics of the radical generation pro-cess, and to test how the latter is influenced by the pres-ence of monomers. It has been shown in previous litera-ture reports that, in these strictly anaerobic conditions, photodegradation of PIs proceeds through formation of photoinduced radical species which evolve, in the ab-sence of scavenging species, to the build-up of photolysis products. Thus, photobleaching rate can be correlated to radical photogeneration efficiency, as recently
exempli-fied by Lalevée et al. or Liska et al.91-95 Kinetics of
chro-mophores photobleaching upon continuous irradiation in the ICT band were investigated upon irradiation with a LED at 365 nm, by measuring the in-time decay of lumi-nescence signal intensity, using thoroughly deaerated (five freeze-pump-thaw cycles with a secondary vacuum pump) solution of V-Shape and Oct in dichloromethane. Evolution of the UV-vis spectra were also monitored in parallel (see Figure 5 and Figure S9).
In the present case, large changes of the isosbestic point at around 360 nm in the absorption spectrum throughout the experiment could be evidenced for both V-Shape and
Oct, with a similar evolution profile in both cases. This
in-dicates a non-straightforward degradation process involv-ing multiple competitive pathways. Two steps at least can be distinguished: in the first 10 minutes, a decrease of ICT band, which can be attributed to the decomposition of
V-9
Shape (or Oct), is associated to the concomitant increase
of the absorbance at 359 nm and 325 nm. Then, while decrease of the ICT band slowly proceeds forward, the newly formed bands at 359 nm and 325 nm in turn start to decrease, leading to a drift of the isosbestic point. This double evolution suggests the formation (first step) and consumption (second step) of degradation products of the PIs.
In spite of the relatively intricate profile of photo-induced degradation process, attempts were made to fit the first ten minutes of the recorded evolutions to a simple kinetic model. A relatively good correlation was obtained in the case of V-Shape and Oct with an apparent first order
re-action (see Figure 6 and Figure S10),96 which seemingly
disagrees with the hypothesis of bimolecular electron transfer. However, one has to take into account that the
range of studied concentrations (ca. 10-6 M), as imposed
by the spectrofluorimetric experiment, may affect the rad-ical generation mechanism: in particular, such low con-centrations are reportedly known to strongly disfavor bi-molecular mechanisms in favor of direct hydrogen ab-straction with the solvent, which might account for the
ob-served discrepancies.97
Photolytic rates were found to follow the trend: BDEBP >>
Oct > V-Shape. This indicates that BDEBP generates
radicals more efficiently than Oct which itself is slightly more efficient than V-Shape. This result is consistent with the fact that the molecular design of V-Shape and Oct
was primarily aimed at improving σTPA rather than the
rad-ical generation efficiency. The difference between
V-Shape and Oct might be attributed to the extra arm of Oct
since the radicals are expected to be formed in α position of the aniline.
Figure 5. Photolysis studies (UV-Vis and fluorescence) of V-Shape in degassed dichloromethane under 365 nm irradiation, 40 mW in absence or presence of Ethyl Acrylate (EA). [V-Shape] = 1.4 x 10-6 mol/L; [EA] = 0 mol/L (left); 1.4 x 10-3 mol/L (center) and 7 x 10-3 mol/L (right).
0 5 10 15 20 25 30 0.0 0.2 0.4 0.6 0.8 1.0 V-Shape Emission Oct Emission BDEBP (Abs at 362 nm) Norm alized Signa l Time (min)
Figure 6. Photolysis of V-Shape (red), Oct (purple) and
BDEBP (green) in degassed dichloromethane under 365 nm
irradiation, 40 mW at isoabsorbance concentration at 365
nm. [V-Shape] = 1.4 x 10-6 mol/L; [Oct] = 8.6 x 10-7 mol/L and [BDEBP] = 2.6 x 10-6 mol/L.
Moreover, it was unambiguously shown that addition of monomer (Ethyl Acrylate EA and DPPHA) markedly slows down the photobleaching of V-Shape and Oct (see Figure 7 and Figure S11). The extent of this slow-down appears to be monomer concentration dependent. Taken together with the electrochemical and Stern-Volmer data (see above), we hypothesize from this last result that the pho-toinitiation mechanism takes place in two consecutive steps i/ an initial photoinduced disproportionation of PI, generating the corresponding radical anion and radical cation pairs; in the absence of monomer or other scav-enging species, these species evolve towards terminal degradation products, ii/ an electron transfer towards (or alternatively, radical addition unto) monomer molecules, which restores at least a part of the chromophore (one or two arms) to its initial state and prevents its evolution to-wards degradation products so that the initial signal of flu-orescence and absorbance is at least partially restored.
10
Figure 7. Photolysis kinetics followed by fluorescence meas-urement of V-Shape (top) and Oct (bottom) in degassed di-chloromethane under 365 nm irradiation, 40 mW in absence or presence of EA. [V-Shape] = 1.4 x 10-6 mol/L; [Oct] = 8.6 x 10-7 mol/L; [EA] = 0 mol/L (blue); 1.4 x 10-3 mol/L (red); 7 x 10-3 mol/L (black).
Two-photon induced 3D microfabrication. With this
in-formation in hands, we finally studied the potential of
V-Shape and Oct for the two-photon polymerization (TPP)
of acrylic monomers. Our main concern was to compare
the polymerization threshold (Pth) obtained with V-Shape
and Oct with benchmark PIs, and see whether lowering of this value could be obtained without affecting the ac-cessible resolution.
As presented earlier, design of the newly synthesized PIs involves hexyl substitution on the terminal aniline moie-ties. The purpose of this long alkyl group was to improve chromophore solubility in the photoresist formulation,
such long π-conjugated structures featuring large σTPA
be-ing particularly prone to aggregation because of π-stack-ing. To take advantage of this feature and maximize PI
solubility, a reactive “diluent” was introduced within the re-sist formulation, in the form of 1,10-decanediol diacrylate (DDA), a difunctionnal acrylic monomer featuring a long fatty alkyl chain. The rest of the resist (80wt%) consisted in dipentaerythritol penta-/hexa-acrylate (DPPHA). The latter was chosen as its high functionality provides an ex-cellent crosslinking density, which improves photoresist sensitivity as well as mechanical strength and reduces
shrinkage of the generated pattern.98
This formulation afforded a maximal 1wt% (10.9 mmol/g) solubility of V-Shape upon direct mixing with the mono-mers. In contrast, the solubility of Oct remained limited to 0.66wt% (5.4 mmol/g) and was not improved with addition of co-solvent and sonication which we attribute to a larger π-conjugated backbone.
In order to evaluate Pth and minimal accessible size of
photogenerated polymers, two types of objects were fab-ricated. We considered that polymerization threshold was reached when the photogenerated structures could up-hold the monomer washing step without undergoing sig-nificant distortion (see Figure 8d). As a first model struc-ture, 20 µm lines were printed directly at the surface of the coverslip paying close attention to the level of the laser focus position in order to be exactly at or slightly above the substrate/photoresist interface and thus measuring the correct linewidth while maintaining a good adhesion to the substrate. Figure 8d is a SEM image of periodic lines fabricated with decreasing power at 40 µm/s writing speed. Although suspended lines are often preferred in
literature for related studies,99-100 this choice of structures
present the advantage of being extremely rapid to imple-ment, on the one hand, and to reduce the impact of the shrinkage on the measurement of the linewidth on the other (structures being stabilized by anchoring to the
co-verslip).101 Moreover, it limits the potential deformation
due to spherical aberrations.15
Figure 8. SEM pictures of representative structures generated by TPP for the measurements of the polymerization thresholds, the linewidths and the line heights. Scan speed: 40 µm/s. (a) Grazing view of a suspended line, photoresist:
DPPHA:DDA:V-Shape (79.2:19.8:1), Pav 103 µW (b) 45° view of suspended lines, photoresist: DPPHA:DDA:V-Shape (79.2:19.8:1), Pav from 257 to 82 µW (c) Top view of a line written at the surface of the coverslip, photoresist: DPPHA:DDA:Oct (79.5:19.8:0.7), Pav 96 µW (d) Top view of lines written at the surface of the coverslip, the white arrow points to the line at the threshold, photoresist: DPPHA:DDA:Oct (79.5:19.8:0.7), Pav from 303 to 85 µW. 0 20 40 60 80 100 0.0 0.2 0.4 0.6 0.8 1.0 Ratio [V-Shape]:[EA] 1:0 1:1000 1:5000 Norm alized Emission Time (min)
11
However, it does not provide reliable measurement of the line heights. Thus for comparison, suspended lines were also printed to check for compliance with the results ob-tained with attached lines, and enable estimation of the aforementioned parameter. 13 µm lines were hanged be-tween 5 µm-separated blocks at 1.5 µm from the surface (see Figure 8a and b) to ensure good mechanical stability. The general protocol used for systematic measurement of line width and height are reported in the experimental sec-tion and detailed in ESI (see Figure S12). Basically, a Py-thon script was designed to provide statistical treatment on a large number of measurements. Generally speaking, no significant difference was found between suspended and surface grafted lines for what concerns linewidth, while the line height and aspect ratio could be measured on suspended lines exclusively, with a value of ca. 2.7 ,
typical of TPP using high NA objective.102 (see Figure
S13). However, it could be noted that the width of the sus-pended lines is slightly inferior to the width of the equiva-lent lines attached to the coverslip at low laser power (near threshold). This could be due to moderate shrinkage of the structure, more visible at low monomer conversion. In the following, we thus chose to focus on surface grafted
lines for the determination of Pth.
Table 2. Comparison of the obtained Pth between the studied photoresists. Composition: DPPHA:DDA (80:20) with 5.4 µmol PI/g; scanning speed: 40 µm/s.
PI TPP Pth (µW) V-Shape 120 Oct 90 BDEBP 550 ITX 690 OXE2 480
Following this procedure we compared Pth between
V-Shape, Oct and our three benchmark compounds, in the
monomer mixture described above (see Table 2). It clearly appears that the polymerization thresholds are significantly improved when comparing Oct and V-Shape to the benchmarks. Threshold is lowered by approxi-mately a factor 5-6 as compared to BDEBP, 6-8 as com-pared to ITX and 4-5 as comcom-pared to OXE2 for V-Shape and Oct respectively. This improvement could be
ex-plained by the much increased σTPA of the newly designed
PIs as compared to commercially available benchmarks.
It clearly indicates that V-Shape and Oct constitute a very promising basis for the future development of improved PIs for two-photon induced polymerization.
In spite of a σTPA more than one order of magnitude higher
for V-shape and Oct than for BDEBP, Pth is only improved
by a factor of ca. 5. The lower absorption of BDEBP seems partly compensated by its better radical generation ability (vide supra). A similar reason might also account for the lower threshold of Oct as compared with V-Shape,
while σTPA measurements would suggest an opposite
trend.
In a last set of experiments, dependence of the accessible structure size to irradiance was evaluated for each com-pound.
As expected, the linewidth increases upon increasing the laser power used during the microfabrication process. Alt-hough the observed evolution looks at first sight to be very dependent on the nature of the PI, the impression van-ishes when the evolution is plotted not versus the absolute laser beam intensity, but versus the ratio of applied power
in regard with the Pth value (see inset in Figure 9 and
Fig-ure S14). Using this representation, it becomes apparent that the photoresists display a power-dependence pattern that is similar for all PIs: in all cases, initial evolution of the structure dimension can be modeled using the “beam shape dependence function” presented below (see
Equa-tion 3) for power that does not exceed twice Pth.30, 103
𝑣𝑖= 𝑤𝑥𝑦√2𝑙𝑛 𝑃𝑖
𝑃𝑡ℎ (3)
Where 𝑣𝑖 is the voxel width obtained with the laser power
𝑃𝑖, 𝑤𝑥𝑦 is the beam waist in the x-y plane, and 𝑃𝑡ℎ is the
TPP threshold. As shown in Figures S15 and 16, the ob-tained linewidths are satisfactorily reproduced with Equa-tion 3.
Figure 9. Linewidth versus average laser power and ratio
average power/threshold power (inset) for the five studied photoresists.
Finally, in order to illustrate the potential of the newly syn-thesized photoresists for low-threshold two-photon print-ing of sub-100 nm resolution structures, we envisioned the fabrication of a 250 nm periodic grating and cubic woodpiles as those presented in Figure 10. A lateral res-olution as small as 80 nm and an axial resres-olution as small as 190 nm were thus demonstrated using a low-cost sub-nanosecond pulsed 532 nm laser (see Figure 8a, c and Figure 10), proving that excellent resolution is maintained with our system.
0 200 400 600 800 1000 1200 1400 1600 1800 2000 0 100 200 300 400 500 600 700 V-Shape Oct BDEBP ITX OXE2 Line width (n m) P (µW) 1 2 3 4 5 0 100 200 300 400 500 600 Linewi dth (nm) P/Pth
12
Figure 10. SEM pictures of representative structures generated by TPP. Writing speed 40 µm/s, photoresist:
DPPHA:DDA:V-Shape (79.2:19.8:1) (a) Top view of a 250 nm periodic grating with a close-up in the upper right corner, Pav 103 µW (b) 45° view of a 700 nm-cubic woodpile, Pav 125 µW (c) Grazing view of a 700 nm-cubic woodpile, Pav 125 µW
CONCLUSIONS
In conclusion, we show that highly efficient radical two-photon initiators can be obtained taking advantage of near-resonant enhancement of the two-photon cross-sec-tion efficiency. Taking a simple π-conjugated structure, that displays intense one-photon absorption in the blue part of the UV-vis spectrum, yet only moderate
two-pho-ton efficiency in the regular TPA range (with λTPA = 2λOPA,,
ca. 800 nm), we show that it is possible to obtain σTPA
val-ues that reach, for one of the studied compounds, 1500 GM at the 532 nm operating wavelength of the laser
(close to λOPA) used in our microfabrication setup.
Owing to their extended π-conjugated structures, the as designed PIs have accessible electrochemical potential in oxidation as well as in reduction, rendering photoinduced disproportionation thermodynamically allowed. As a re-sult, photoinduced radical generation takes place upon PI irradiation, which can be used to efficiently initiate polymerization reactions. Both PIs being functionalized by long alkyl chain that favors their solubility in organic me-dia, they can be readily solubilized within acrylic mixtures for direct use in two-photon microfabrication without addi-tional preparation step. We illustrate the benefit of this ap-proach by comparing the polymerization threshold ob-tained with both newly synthesized PIs to Norrish I and Norrish II type ones classically used for two-photon micro-fabrication with sub-nanosecond pulsed 532 nm lasers. Using similar photoresist formulation, the polymerization threshold was lowered up to eight times with the reported new PIs as compared to benchmarks, similar to some of the best reported PIs to date, in spite of a comparatively lower radical generation efficiency. Finally, we show that the improvement in polymerization threshold is not ob-tained at the expense of resolution, as illustrated with the-fabrication of thin bridges and periodic structures with pe-riodic distance of 250 nm which makes both new PIs
(V-Shape and Oct) excellent candidates for massive parallel
printing of large surfaces.
Funding Sources
C.A. thanks the Agence Nationale de la Recherche ANR for financial support (PhD grant) in the framework of the “New3DPrint” project (ANR-17-CE08-0037). This work has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement No 780278. The dissemination of results herein reflects only the author's view and the European Commission is not re-sponsible for any use that may be made of the information it contains.
Notes
The authors declare no competing financial interest. Patent Pending, Registry number FR2007303 ACKNOWLEDGMENT
The authors thank the Centre Régional de Mesures Phy-siques de l'Ouest for the high resolution mass spectrom-etry measurements and the Pôle Scientifique de Mo-délisation Numérique/Institut Universitaire de France (PSMN/IUF) for computational resources.
REFERENCES
1. Tofail, S. A. M.; Koumoulos, E. P.; Bandyopadhyay, A.; Bose, S.; O’Donoghue, L.; Charitidis, C., Additive manufacturing: scientific and technological challenges, market uptake and opportunities. Materials Today 2018, 21 (1), 22-37.
2. Vaidya, S.; Ambad, P.; Bhosle, S., Industry 4.0 – A Glimpse. Procedia Manufacturing 2018, 20, 233-238.
3. Ligon, S. C.; Liska, R.; Stampfl, J.; Gurr, M.; Mülhaupt, R., Polymers for 3D Printing and Customized Additive Manufacturing. Chemical Reviews 2017, 117 (15), 10212-10290.
4. Melchels, F. P. W.; Domingos, M. A. N.; Klein, T. J.; Malda, J.; Bartolo, P. J.; Hutmacher, D. W., Additive manufacturing of tissues and organs. Progress in Polymer Science 2012, 37 (8), 1079-1104.
5. Narayan, R. J.; Doraiswamy, A.; Chrisey, D. B.; Chichkov, B. N., Medical prototyping using two photon polymerization. Materials Today 2010, 13 (12), 42-48.
13
6. Bos, F.; Wolfs, R.; Ahmed, Z.; Salet, T., Additive manufacturing of concrete in construction: potentials and challenges of 3D concrete printing. Virtual and Physical Prototyping 2016, 11 (3), 209-225.
7. Mehrpouya, M.; Dehghanghadikolaei, A.; Fotovvati, B.; Vosooghnia, A.; Emamian, S. S.; Gisario, A., The Potential of Additive Manufacturing in the Smart Factory Industrial 4.0: A Review. Applied Sciences 2019, 9 (18), 3865.
8. Attaran, M., The rise of 3-D printing: The advantages of additive manufacturing over traditional manufacturing. Business Horizons 2017, 60 (5), 677-688.
9. Bikas, H.; Stavropoulos, P.; Chryssolouris, G., Additive manufacturing methods and modelling approaches: a critical review. The International Journal of Advanced Manufacturing Technology 2016, 83 (1), 389-405.
10. Chatham, C. A.; Long, T. E.; Williams, C. B., A review of the process physics and material screening methods for polymer powder bed fusion additive manufacturing. Progress in Polymer Science 2019, 93, 68-95.
11. Lim, M. P.; Guo, X.; Grunblatt, E. L.; Clifton, G. M.; Gonzalez, A. N.; LaFratta, C. N., Augmenting mask-based lithography with direct laser writing to increase resolution and speed. Opt. Express 2018, 26 (6), 7085-7090.
12. Stansbury, J. W.; Idacavage, M. J., 3D printing with polymers: Challenges among expanding options and opportunities. Dental Materials 2016, 32 (1), 54-64.
13. Maruo, S.; Ikuta, K., Submicron stereolithography for the production of freely movable mechanisms by using single-photon polymerization. Sensors and Actuators A: Physical 2002, 100 (1), 70-76.
14. Park, S.-H.; Yang, D.-Y.; Lee, K.-S., Two-photon stereolithography for realizing ultraprecise three-dimensional nano/microdevices. Laser & Photonics Reviews 2009, 3 (1‐2), 1-11.
15. Burmeister, F.; Steenhusen, S.; Houbertz, R.; Zeitner, U. D.; Nolte, S.; Tünnermann, A., Materials and technologies for fabrication of three-dimensional microstructures with sub-100 nm feature sizes by two-photon polymerization. Journal of Laser Applications 2012, 24 (4), 042014.
16. Pearre, B. W.; Michas, C.; Tsang, J.-M.; Gardner, T. J.; Otchy, T. M., Fast micron-scale 3D printing with a resonant-scanning two-photon microscope. Additive Manufacturing
2019, 30, 100887.
17. Farsari, M.; Chichkov, B. N., Two-photon fabrication. Nature Photonics 2009, 3 (8), 450-452.
18. Poocza, L.; Gottschaldt, M.; Markweg, E.; Hauptmann, N.; Hildebrand, G.; Pretzel, D.; Hartlieb, M.; Reichardt, C.; Kübel, J.; Schubert, U. S.; Mollenhauer, O.; Dietzek, B.; Liefeith, K., Optimized Photoinitiator for Fast Two-Photon Absorption Polymerization of Polyester-Macromers for Tissue Engineering Advanced Engineering Materials 2017, 19 (3), 1600686.
19. Maruo, S.; Fourkas, J. T., Recent progress in multiphoton microfabrication. Laser & Photonics Reviews
2008, 2 (1‐2), 100-111.
20. Houbertz, R.; Fröhlich, L.; Popall, M.; Streppel, U.; Dannberg, P.; Bräuer, A.; Serbin, J.; Chichkov, B. N., Inorganic– Organic Hybrid Polymers for Information Technology: from Planar Technology to 3D Nanostructures. Advanced Engineering Materials 2003, 5 (8), 551-555.
21. Stender, B.; Mantei, W.; Houbertz, R., From Lab to Fab — High-Precision 3D Printing. Laser Technik Journal 2017, 14 (2), 20-23.
22. Stender, B.; Hilbert, F.; Dupuis, Y.; Krupp, A.; Mantei, W.; Houbertz, R., Manufacturing strategies for scalable high-precision 3D printing of structures from the micro to the macro range. In Advanced Optical Technologies, 2019; Vol. 8, p 225. 23. Houbertz, R.; Steenhusen, S.; Stichel, T.; Sextl, G., Two-Photon Polymerization of Inorganic-Organic Hybrid Polymers as Scalable Technology Using Ultra-Short Laser Pulses. In Coherence and Ultrashort Pulse Laser Emission Duarte, F. J., Ed. InTech: 2010.
24. Houbertz, R.; Satzinger, V.; Schmid, V.; Leeb, W.; Langer, G., Optoelectronic printed circuit board: 3D structures written by two-photon absorption. SPIE: 2008; Vol. 7053. 25. Xu, B.; Du, W.-Q.; Li, J.-W.; Hu, Y.-L.; Yang, L.; Zhang, C.-C.; Li, G.-Q.; Lao, Z.-X.; Ni, J.-C.-C.; Chu, J.-R.; Wu, D.; Liu, S.-L.; Sugioka, K., High efficiency integration of three-dimensional functional microdevices inside a microfluidic chip by using femtosecond laser multifoci parallel microfabrication. Scientific Reports 2016, 6 (1), 19989.
26. Geng, Q.; Wang, D.; Chen, P.; Chen, S.-C., Ultrafast multi-focus 3-D nano-fabrication based on two-photon polymerization. Nature Communications 2019, 10 (1), 2179. 27. Anseth, K. S.; Bowman, C. N.; Peppas, N. A., Polymerization kinetics and volume relaxation behavior of photopolymerized multifunctional monomers producing highly crosslinked networks. Journal of Polymer Science Part A: Polymer Chemistry 1994, 32 (1), 139-147.
28. Kuebler, S. M.; Rumi, M.; Watanabe, T.; Braun, K.; Cumpston, B. H.; Heikal, A. A.; Erskine, L. L.; Thayumanavan, S.; Barlow, S.; Marder, S. R.; Perry, J. W., Optimizing Two-Photon Initiators and Exposure Conditions for Three-Dimensional Lithographic Microfabrication. Journal of Photopolymer Science and Technology 2001, 14 (4), 657-668.
29. Ligon, S. C.; Husár, B.; Wutzel, H.; Holman, R.; Liska, R., Strategies to Reduce Oxygen Inhibition in Photoinduced Polymerization. Chemical Reviews 2014, 114 (1), 557-589. 30. Sun, H.-B.; Takada, K.; Kim, M.-S.; Lee, K.-S.; Kawata, S., Scaling laws of voxels in two-photon photopolymerization nanofabrication. Applied Physics Letters 2003, 83 (6), 1104-1106.
31. Waller, E. H.; Von Freymann, G., Spatio-Temporal Proximity Characteristics in 3D μ-Printing via Multi-Photon Absorption. Polymers 2016, 8 (8), 297.
32. Schafer, K. J.; Hales, J. M.; Balu, M.; Belfield, K. D.; Van Stryland, E. W.; Hagan, D. J., Two-photon absorption cross-sections of common photoinitiators. Journal of Photochemistry and Photobiology A: Chemistry 2004, 162 (2), 497-502. 33. Cumpston, B. H.; Ananthavel, S. P.; Barlow, S.; Dyer, D. L.; Ehrlich, J. E.; Erskine, L. L.; Heikal, A. A.; Kuebler, S. M.; Lee, I. Y. S.; McCord-Maughon, D.; Qin, J.; Röckel, H.; Rumi, M.; Wu, X.-L.; Marder, S. R.; Perry, J. W., Two-photon polymerization initiators for three-dimensional optical data storage and microfabrication. Nature 1999, 398 (6722), 51-54.
34. Fouassier, J.-P.; Morlet-Savary, F.; Lalevée, J.; Allonas, X.; Ley, C., Dyes as Photoinitiators or Photosensitizers of Polymerization Reactions. Materials 2010, 3 (12), 5130-5142. 35. Lee, K.-S.; Kim, R. H.; Yang, D.-Y.; Park, S. H., Advances in 3D nano/microfabrication using two-photon initiated polymerization. Progress in Polymer Science 2008, 33 (6), 631-681.
36. Zhou, R.; Malval, J.-P.; Jin, M.; Spangenberg, A.; Pan, H.; Wan, D.; Morlet-Savary, F.; Knopf, S., A two-photon active chevron-shaped type I photoinitiator designed for 3D stereolithography. Chemical Communications 2019, 55 (44), 6233-6236.
